BMC Evolutionary Biology BioMed Central

Research article Open Access Phylogenetic relationships and biogeography of the Algansea Girard (: ) of central Mexico inferred from molecular data Rodolfo Pérez-Rodríguez*1,2, Omar Domínguez-Domínguez2, Gerardo Pérez Ponce de León3 and Ignacio Doadrio4

Address: 1Posgrado en Ciencias Biológicas, Instituto de Biología, Universidad Nacional Autónoma de México, México, D.F., México, 2Laboratorio de Biología Acuática, Facultad de Biología, Universidad Michoacana de San Nicolás de Hidalgo, Morelia, Michoacán, México, 3Instituto de Biología, Universidad Nacional Autónoma de México, Departamento de Zoología. Ap. Postal 70-153, C.P. 04510 México D.F., México and 4Departamento de Biodiversidad y Biología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, c/José Gutiérrez Abascal 2, E-28006 Madrid, España Email: Rodolfo Pérez-Rodríguez* - [email protected]; Omar Domínguez-Domínguez - [email protected]; Gerardo Pérez Ponce de León - [email protected]; Ignacio Doadrio - [email protected] * Corresponding author

Published: 7 September 2009 Received: 15 January 2009 Accepted: 7 September 2009 BMC Evolutionary Biology 2009, 9:223 doi:10.1186/1471-2148-9-223 This article is available from: http://www.biomedcentral.com/1471-2148/9/223 © 2009 Pérez-Rodríguez et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract Background: The genus Algansea is one of the most representative freshwater fish groups in central Mexico due to its wide geographic distribution and unusual level of endemicity. Despite the small number of species, this genus has had an unsettled taxonomic history due to high levels of intraspecific morphological variation. Moreover, several phylogenetic hypotheses among congeners have been proposed but have had the following shortcomings: the use of homoplasious morphological characters, the use of character codification and polarisation methods that lacked objectivity, and incomplete taxonomic sampling. In this study, a phylogenetic analysis among species of Algansea is presented. This analysis is based upon two molecular markers, the mitochondrial gene cytochrome b and the first intron of the ribosomal protein S7 gene. Results: Bayesian analysis based on a combined matrix (cytochrome b and first intron S7) showed that Algansea is a monophyletic group and that Agosia chrysogaster is the sister group. Divergence times dated the origin of the genus around 16.6 MYA, with subsequent cladogenetic events occurring between 6.4 and 2.8 MYA. When mapped onto the molecular phylogenetic hypothesis, the character states of three morphological characters did not support previous hypotheses on the evolution of morphological traits in the genus Algansea, whereas the character states of the remaining six characters partially corroborated those hypotheses. Conclusion: Monophyly of the genus Algansea was corroborated in this study. Tree topology shows the genus consists of three main lineages: Central-Eastern, Western, and Southern clades. However, the relationships among these clades remained unresolved. Congruence found between the available geological and climatic history and the divergence times made it possible to infer the biogeographical history of Algansea, which suggested that vicariance events were responsible for the evolutionary history of the genus. Interestingly, this pattern was shared with other members of the freshwater fish fauna of central Mexico. In addition, molecular data also show that some morphological traits alleged to represent synapomorphies in previous studies were actually homoplasies. Others traits were corroborated as synapomorphies, particularly in those species of a subgroup corresponding with the Central-Eastern clade within Algansea; this corroboration is interpreted as a result of evolutionary adaptations.

Page 1 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

Background cal characters [9,11,12,23]. Such molecular approaches Mexico lies between the Nearctic and Neotropical biogeo- demonstrated that fish diversity was in fact underesti- graphical zones and is considered to be a transitional mated because several new species were described; in zone. Because of this, it is possible to find fauna with dif- addition, some morphological characters commonly used ferent evolutionary origins [1]. One of the world's great in the classification of these groups were homoplasies. tropical-subtropical highlands is the massive uplift The genus Algansea is one of these groups that exemplify known as the Mesa Central of Mexico (MCM) and its the particular problem of using morphological characters southern limit, the Trans Mexican Volcanic Belt (TMVB). as the only source of information to establish a phyloge- Since the Miocene, this zone has experienced an active netic hypothesis. Although this genus possesses a few rep- geological history, which has promoted a complex surface resentative species, Algansea aphanea, Algansea avia, configuration, including a wide variety of ecosystems. The Algansea barbata, Algansea lacustris, Algansea monticola, freshwater fish fauna of the MCM [2,3] is unique, with Algansea popoche, Algansea tincella, [5], and Algansea ame- around 78 species, and is represented by an unusual level cae [24], its taxonomic history has been unstable, particu- of endemicity (70%) [4,5]. Most of the endemic species larly due to high levels of intraspecific morphological are represented by monophyletic groups that have under- variability found in the most widely distributed species gone a diversification process within central Mexico [6,7] within the genus, A. tincella [21]. such as the entire subfamily (41 species), the Atherinopsid genera Chirostoma (19 species), and three The first significant attempt to elucidate the systematic endemic genera belonging to the family Cyprinidae: relationships among species of Algansea was made by Bar- Algansea (7 species), Evarra (3 species), and Yuriria (3 spe- bour and Miller [21]. These authors studied only six valid cies) [6,8-12]. species and two subspecies and found two groups: (1) species with maxillary barbels (A. monticola - including A. Knowledge about the diversification processes of freshwa- m. monticola and A. m. avia -, A. barbata and A. aphanea) ter fish in central Mexico, including their origin as well as and (2) species without maxillary barbels (A. lacustris, A. their evolutionary and biogeographical history, is still popoche and A. tincella). In addition, based on the presence incomplete. Several hypotheses regarding the biogeogra- of a plate-like dermosphenotic bone, the genus Gila was phy of freshwater fish in the region have been discussed in determined to be the closest relative to Algansea [21]. some detail in several studies [2,3,13-17]. These authors However, current molecular studies of Mexican cyprinids described general patterns using occurrence data and mor- do not support this relationship, and these analyses estab- phological comparisons. More recently, studies that incor- lished the genus Agosia as the sister group of Algansea porated molecular approaches in a phylogenetic context [10,12]. Subsequent to Barbour and Miller's hypothesis, have been conducted to elucidate the biogeographical and the subspecies A. m. avia was validated as an independent evolutionary history of fish in central Mexico among species, A. avia. This species, together with A. monticola, A. groups such as poecilids [18], goodeids [7,9,19], and barbata, and A. aphanea, are the barbeled species. The last cyprinids [10,12]. All of these studies agree that the histor- taxonomic change implied that the highly differentiated ical biogeography of central Mexico and its freshwater fish population of A. tincella from the Ameca River should be fauna is linked to the intense geological activity since the recognised as A. amecae [24]. All the aforementioned sug- early Miocene. This activity has generated a complex gest that the morphological characters on which those hydrological system characterised by high dynamism and first analyses were based have evolved independently sev- the formation and destruction of drainages. This dyna- eral times throughout the evolutionary history of mism promoted vicariance, taxon-pulse, and species- cyprinids [25], resulting in controversial classification pulse events [20]. However, the complexity of these bio- schemes [26]. geographic patterns and the few fish groups studied thus far make it necessary to study other co-distributed fish Molecular markers have proven to be very useful tools groups, such as members of the genus Algansea, to formu- when elucidating the phylogenetic relationships of verte- late a more robust biogeographical scenario of the area. brate groups. In the last few years, the mitochondrial gene cytochrome b (cyt b) and the first intron of the S7 ribos- The first attempts to uncover phylogenetic patterns of omal protein gene have been used in different phyloge- freshwater fish in central Mexico were based on just a few netic analyses to provide insights into the fish morphological traits, resulting in a non-robust hypothesis evolutionary history at various taxonomic levels. Both regarding the evolutionary history of the groups genes have proven their utility not only for inferring the [2,3,21,22]. More recently, phylogenetic studies on differ- phylogenetic relationships among closely related species ent freshwater fish families occurring in the region that but also for investigating intraspecific variation and even were based on various molecular markers revealed results for establishing species boundaries [27-31]. Accordingly, contradictory to those results obtained with morphologi- in this paper we used both cyt b and the first intron of S7

Page 2 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

to infer the phylogenetic relationships among species and quency distribution was always homogeneous among populations of Algansea throughout their entire distribu- taxa. tion range in central Mexico. Additionally, the results of these analyses were used to: 1) compare with the previous Phylogenetic Analysis phylogenetic hypotheses based on morphological data; 2) Details of the Maximum Parsimony (MP) and Bayesian characterise the evolution of the main morphological Inference (BI) analyses are summarised in Table 1. All the characters used for identifying natural groups; and 3) analyses (mitochondrial, nuclear, and combined data) establish possible biogeographic scenarios in which spe- showed almost identical and well-supported topologies; cies and populations may have evolved in the context of however, the MP tree was less resolved than the Bayesian the geological and climatic history and contrast this his- consensus tree. Thus, we focus our discussion on the more tory with other co-distributed taxa. resolved Bayesian tree and only summarise the results of the MP analysis. Results Sequence data patterns All the cladograms retrieved Algansea as monophyletic In the cyt b sequences (1,140 bp), a total of 425 sites with high support values (Figures. 1, 2 and 3). Also, all the (37%) were variable, and 279 (24%) were parsimony species formed monophyletic assemblages, except A. tin- informative. As expected for a protein-coding gene, third cella and A. amecae in the nuclear gene cladogram (Figure codon positions were the most variable (320), followed 2). In addition, all the trees showed Agosia chrysogaster as by the first (74) and the second (31) positions. For the S7 the sister group to Algansea (Figures 1, 2 and 3). intron, the size was 965 bp including gaps, and there were 279 variable sites, 103 of which were parsimony informa- Sister group relationships among species of Algansea were tive. Comparison of the individual genes' phylogenetic generally consistent with the cyt b or combined data sets. performance under taxon sampling of the combined anal- Three main clades corresponding to the Central-Eastern, ysis is found in Table 1. Estimated parameters using Mod- Western, and Southern regions of the whole distribution eltest [32] are shown in Table 1. As in other North range of the genus Algansea were recovered with high sup- American cyprinids' cyt b, a low Guanine frequency in the port values. The Central-Eastern clade is represented by third codon positions was found [12,23,33]. S7 sequences the species A. barbata, A. amecae, A. lacustris, and A. tin- were AT-rich, as has been previously reported for other cella (Figures 1, 2 and 3). The Western clade included the families of freshwater fishes [34,35] and other North two species A. monticola and A. avia from the northwest- American cyprinids [12]. Despite such apparent bias, the ern headwaters of the Santiago River and the lower San- χ2 test for base homogeneity indicated that the base fre- tiago, respectively (Figures 1 and 3). The last clade is

Table 1: Parameters and statistics summary of phylogenetic analyses

cyt b

Codon position 1st 2nd 3rd allpos S7 cyt b + S7

MP analysis Number of sites 380 380 380 1140 965 2105 Number of variables sites 74 31 320 424 279 675 Parsimony-informative sites 34 2 243 279 103 365 Most parsimonious trees - - - 1 1 1 Tree length - - - 879 343 1183 Consistency index - - - 0.622 0.907 0.693 Retention index - - - 0.830 0.819 0.666

BI analyses Model (BIC) TrNef+G HKY TrN+G TrN+G HKY+G - Nucleotide proportions A = 24% A = 20% A = 35% A = 26% A = 31% - C = 25% C = 26% C = 34% C = 29% C = 15% - G = 26% G = 13% G = 11% G = 17% G = 20% - T = 25% T = 40% T = 20% T = 28% T = 34% - χ2 test of base frequencies χ2 = 4.62 χ2 = 1.28 χ2 = 1.28 χ2 = 21.68 χ2 = 3.71 - gl = 153 gl = 153 gl = 153 gl = 153 gl = 54 p = 1.00 p = 1.00 p = 1.00 p = 1.00 p = 1.00 Alpha 0.23 t.i. 2.45 0.21 1.41 -

BIC = Bayesian information

Page 3 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

100/- Verde-Santiago 96/- Middle Lerma -/- 100/- Cuitzeo A. tincella

Zacapu clade Central-Eastern

100/- -/- Panuco

100/- Lower Lerma Verde-Santiago 100/94 100/100 Balsas 98/- San Gregorio -/- A. lacustris 100/100 100/98 Patzcuaro 100/94 San Gregorio 100/92 100/100 Ameca A. amecae -/- upper Lerma A. barbata 100/100

100/98 Santiago Western clade 100/100 A. avia 100/98 Compostela 100/- 96/98 Bolaños-Santiago A. monticola 100/100 100/100

-/- Southern clade Ayuquila-Armería 100/100 A. aphanea Tamazula-Coahuayana A. chrysogaster 100/- D. melanops C. ornatum G. robusta 0.1

FigureThe consensus 1 tree from Bayesian analysis of the cytochrome b sequences The consensus tree from Bayesian analysis of the cytochrome b sequences. The phylogeny reported corresponds to consensus topology of 8,000 trees sampled using Bayesian analysis. Upper numbers correspond to posterior probabilities (val- ues < 95 are not shown), and lower numbers correspond to bootstrap support (values < 90 are not shown).

Page 4 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

Cuitzeo 100/95

97/- Pánuco Central-Eastern clade Central-Eastern Ameca A. tincella 97/-

Verde-Santiago 100/94

100/97 Balsas

Patzcuaro A. lacustris

Ameca A. amecae

upper Lerma A. barbata 100/-

Bolaños-Santiago A. monticola Western clade 100/100

100/- Santiago

100/97 A. avia

Compostela

100/98 Southern clade Ayuquila-Armeria 100/100 A. aphanea

-/- Tamazula-Coahuayana

A. chrysogaster

C. ornatum

D. melanops

G. robusta 0.1

TheFigure consensus 2 tree from Bayesian analysis of the first intron of the S7 sequences The consensus tree from Bayesian analysis of the first intron of the S7 sequences. The phylogeny reported corre- sponds to consensus topology of 9,750 trees sampled from Bayesian analysis. Upper numbers correspond to posterior proba- bilities (values < 95 are not shown), and lower numbers correspond to bootstrap support (values < 90 are not shown).

Page 5 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

Cuitzeo 100/92 Panuco 100/-

A. tincella clade Central-Eastern Verde-Santiago 100/- 100/95 Balsas

100/- Patzcuaro A. lacustris

100/- 100/100 Ameca A. amecae

upper Lerma A. barbata

100/100 Western clade 100/100 Santiago 100/100 A. avia

100/- Compostela

Bolaños-Santiago A. monticola 100/100 -/- Southern clade

Ayuquila-Armeria 100/100 A. aphanea

Tamazula-Coahuayana

A. chrysogaster

D. melanops -/- G. robusta

C. ornatum 0.1

ConsensusFigure 3 tree from Bayesian analysis of the combined cytocrome b + S7 intron data sets Consensus tree from Bayesian analysis of the combined cytocrome b + S7 intron data sets. The phylogeny reported corresponds to consensus topology of 9,900 trees sampled from Bayesian analysis. Upper numbers correspond to posterior probabilities (values < 95 are not shown), and lower numbers correspond to bootstrap support (values < 90 are not shown).

Page 6 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

represented only by A. aphanea (Figures 1, 2 and 3), a spe- to D = 0.008 ± 0.004. The highest divergences were cies that inhabits the Southern Pacific river basins of Ayu- p quila-Armería and Coahuayana-Tamazula (see Additional found between A. aphanea and A. monticola when com- File 1 and Figure 4). The basal relationships among the pared to the remaining species of Algansea, which had val- three main clades were not resolved. ues ranging from Dp = 0.020 ± 0.005 to Dp = 0.036 ± When intron S7 was used, there were no differences 0.006. between both phylogenetic methods (MP and BI); inclu- sion or exclusion of indels in the MP analyses gave the Molecular clock estimates among the genera Algansea and same topology. However, with respect to the cyt b and the Agosia placed the time to the most recent common ances- combined data sets, a lower resolution tree was formed tor (TMRCA) at 14.7 MYA (16.3-10.6 MYA). The esti- with a resulting basal polytomy (Figure 2). mated age of the diversification of the Central-Eastern, Western, and Southern clades is 6.2 MYA (7.8-4.7 MYA). The main difference between the three data sets was the The split between A. avia and A. monticola occurred 4.5 position of the species A. barbata with respect to all the MYA (6.2-3.2 MYA). Within the Central-Eastern clade, the congeners. In the cyt b and combined data analyses, A. bar- oldest splitting event occurred about 2.9 MYA (4.1-2.1 bata is included in the Central-Eastern clade (Figures 1 MYA) between A. barbata and the clade containing the and 3). However, in the S7 analysis, A. barbata is associ- ancestor of the A. amecae, A. lacustris, A. tincella. Following ated with a member of the Western clade (Figure 2). The the cladogenetic events, A. amecae and the clade A. lacus- most resolved and best-supported hypothesis was tris, A. tincella were dated to 2.4 MYA (3.1-1.9 MYA), and obtained with combined Bayesian analyses (Figure 3). A. lacustris and A. tincella were dated to 1.9 MYA (2.6-1.3 MYA). For cyt b, the genetic divergences found among species Mapping some morphological traits commonly used in ranged from Dp = 0.020 ± 0.004 found between A. tincella the classification of Algansea onto the molecular phyloge- and A. lacustris to 0.090 ± 0.008 found between A. aphanea nies clearly shows that the orientation of the mouth, and A. avia. Genetic divergences among species of the either terminal or upturned, represents a homoplasious character (Figure 5A). Except for a subgroup of the Cen- Central-Eastern clade showed relatively low values, with tral-Eastern clade formed by A. amecae, A. lacustris, and A. the highest value (Dp = 0.038 ± 0.006) observed between tincella, the supraethmoid orientation and the gut flexure A. lacustris and A. amecae and the lowest value (Dp = 0.020 were homoplasies as well (Figure 5D, F). The lack of max- ± 0.004) observed between A. lacustris and A. tincella. illary barbels, a higher number of gill rakers (Figure 5B), Divergences between the species of the Western clade an irregular supraethmoid margin, and well-developed epiotic bones (Figure 5C) were synapomorphies for the accounted for the higher values, with D = 0.078 ± 0.008. p aforementioned subgroup. In addition, the remaining Among the three main clades, the observed divergence two characters mapped, i.e., standard length (as was values were similar: D = 0.083 ± 0.008 between the coded in the present study) and the neurocranium dome, p were synapomorphies for the whole genus, supporting the

Southern and Central-Eastern clades, followed by Dp = monophyly of the Central-Eastern, Western and Southern clades (Figure 5E, G). 0.086 ± 0.007 between the Western and Central-Eastern clades, and Dp = 0.088 ± 0.007 between the Southern and Barbour and Miller's [21] and Jensen and Barbour's [36] Western clades. Within species, the lowest divergence analyses demonstrated that the standard length was not useful for establishing sister-group relationships among value was Dp = 0.002 ± 0.001 between both analysed species of Algansea. In our study, this character was re-eval- populations of A. aphanea. The highest divergences within uated and was found to be informative with the proper species were found in populations of A. tincella, particu- codification, i.e., smaller (< 100 mm) or larger size (> 100 mm). Standard length character was found to be a synapo- larly between the population from the Quitupan River morphy for a species group formed by Southern and West- (Balsas Basin), which had a divergence value of Dp = ern clades (smaller group) and the Central-Eastern clade 0.014-0.018 ± 0.004 with respect to the remaining A. tin- (larger group) (Figure 5E). cella populations. For the S7 intron, the lowest diver- Discussion gences between species were found between A. tincella and Phylogenetic relationships A. lacustris, with values ranging from Dp = 0.005 ± 0.002 The monophyly of the genus Algansea was corroborated using molecular markers. The position of the monotypic

Page 7 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

FigureLocalities, 4 cladogenetic, and geologic events in the evolutionary and biogeographic history of Algansea Localities, cladogenetic, and geologic events in the evolutionary and biogeographic history of Algansea. Paleo- lakes: Az = Aztlan; CH = Chapala (Gray area inside the paleolake indicates the current Chapala lake extension); Ch = Chincua; Ma = Maravatio; Za = Zacoalco. Faults and grabens: Am = Ameca; SM = San Marcos. TCVF = Tarascan Corridor Volcanic Field; SJ = San Juanico lake.

Page 8 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

AncestralFigure 5 character-state mapping on the combined hypothesis of Algansea Ancestral character-state mapping on the combined hypothesis of Algansea. A) 1. orientation of mouth; B) 2. maxil- lary barbels and 3. gill rakers; C) 4. supraetmoid anterior margin and 5. Epiotic bones; D) 6. supraethmoid orientation; E) 7. standard length; F) 8. gut flexure; and G) 9. Neurocranium dome. All character-state codes were based on Barbour and Miller (1978), except supraethmoid orientation, which was based on Jensen and Barbour (1981), and standard length, which was recoded in the present study.

Page 9 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

genus Agosia as a sister group of Algansea was consistent in fish species suggests that the process of duplication of the all analyses. The relationship of the genus Gila with Algan- S7 gene is considerably lower in fish than in mammals sea as previously suggested by Barbour and Miller [21], [43]. In addition, no duplication cases have been found in was rejected. Our results support previous studies in previous studies on fish where the S7 intron was used which a close phylogenetic relationship was found [12,28,30,31,35,38,42-44]; therefore, a duplication event between Algansea and Agosia [10,12,37]. probably did not occur within Algansea.

Sequence variation for the mitochondrial cyt b and Distinguishing between hybridisation and deep coales- nuclear S7, independently or combined, did not provide cence is complicated because both processes generate sim- enough information to resolve the relationships among ilar phylogenetic patterns [40,41,45,46]. Although the three well-supported groups: the Central-Eastern, introgressive hybridisation have played an important role Western, and Southern clades. A lower supported relation- in the evolutionary history of some North American ship, Western clade + Southern clade, was found when the cyprinids [47-50], Barbour and Miller [21] recognised cyt b and the combined matrix were used (Figures 1 and hybrids in a zone where A.tincella and A. popoche are sym- 3), whereas a basal polytomy was found using the S7 patric. There are two reasons for discarding a recent and/ intron (Figure 2). Phylogenetic reconstruction shows no or historical hybridisation as the cause of incongruence resolution in basal members of Algansea, which might be between mitochondrial and nuclear genes-based hypoth- the result of distinct evolution rates of the markers used in eses (due to the unstable position of A. barbata). On the this study. The evolution rate of cyt b was almost twice as one hand, disjunctive historical and current distribution high as the evolution rate of nuclear S7 intron. This fact patterns of A. barbata and A. monticola (see Biogeographic had been previously reported for other fish [30,31,38]. Implications and Figure 4) suggest that no contact However, Schönhuth et al. [12] recently showed that, between the two species occurred. Second, as in other although the evolution rates of cyt b and S7 differ mark- North American cyprinids [48-50] and other fish groups edly, both markers had a similar performance in recon- [51,52], when a hybridisation event is detected by phylo- structing the phylogenetic relationships among species genetic analysis, it needs further corroboration using data and genera of North American cyprinids. from different sources. However, data gathered thus far on the evolutionary history of Algansea show contradicting Therefore, this lack of resolution, especially at the base of results with respect to the relationships among A. barbata the cladogram of the genus, as shown by separate and and other congeners. For instance, morphological charac- combined analyses, may be explained as a result of inad- ters support the relationships between A. barbata and A. equate sampling of data, sampling of taxa, or a composi- aphanea [36], while topology obtained through the cyt b tional base bias. However, such phylogenetic pattern gene placed A. barbata within the Central-Eastern clade as could be caused by a process implicit to the species' evo- the sister species of the not-barbeled species (Figure 1). lutionary history as well, such as a rapid or simultaneous The topology obtained through the first S7 intron shows speciation [39]. that A. barbata and A. monticola are sister taxa (Figure 2).

The combined data analysis provided the best resolution Based on the premise that the S7 intron shows a slower for establishing the phylogenetic relationships among evolution rate, it is possible that inconsistencies may members of Algansea, and that analysis also shows the reflect insufficient time to complete the lineage sorting. highest support values. Particularly for the Central-East- An incomplete lineage sorting is found in relatively recent ern clade, the combined analysis represented the best esti- diversification events, and in particular when nuclear mate for the relationships among the included species genes are employed a longer time is required to reach (Figure 3). reciprocal monophyly [53]. Given the rapid speciation of the genus Algansea in central Mexico, relationship Otherwise, the inconsistent phylogenetic position of A. between A. barbata and A. monticola obtained in the S7 barbata, in particular in the mitochondrial and nuclear intron topology is attributed to the retention of an ances- trees, is the main difference found between topologies tral polymorphism. In fact, incongruence was observed obtained by using mitochondrial and nuclear genes. The between the three kinds of evidence, morphological, and lack of congruence among molecular markers could be the mitochondrial and nuclear genes, reflecting the ran- explained by gene duplication, hybridisation (ancient or dom nature of lineage sorting [40,46]. recent), or incomplete lineage sorting [40,41]. Biogeographic implications Despite the existence of pseudogenes in the S7 ribosomal Based on the sister group relationships between Algansea protein gene in mice [42], the amplification of single PCR and Gila as found by Barbour and Miller [21], it was pos- products for the first and second introns among distant tulated that the genus Algansea derived from a widespread

Page 10 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

ancestor in the highlands of western Mexico during the The event that originated the separation between A. bar- late Tertiary. Our findings indicate that the sister group of bata and the ancestor of the clade formed by A. tincella, A. Algansea is the genus Agosia. This genus is distributed from lacustris, and A. amecae is calculated to have occurred 2.9 the Western Sierra Madre Occidental in Mexico to the MYA (4.1-2.1 MYA), reinforcing the idea that climatic and Rocky Mountains in the United States (Figure 4), support- geological changes in the MCM during that time led to the ing a western origin and a colonisation route to central diversification of Algansea. During this period, intense Mexico by the ancestor of the genus Algansea. tecto-volcanic activity combined with a dry period pro- duced a strong compartmentalisation of palaeolakes in According to the estimated divergence times, the cladog- central Mexico. This activity could have been responsible enetic event involving the ancestor of Agosia and Algansea for population isolation and subsequent speciation events was dated about 14.7 MYA (16.3-10.6 MYA), during the [56,59]. Likewise, this distributional pattern is in agree- middle Miocene (Figure 6). This event is associated with ment with the separation of two lineages of the goodeid the uplift of the western part of the MCM and the south- tribe Girardinichthyini that occurred in the palaeolakes of ern part of the Sierra Madre Occidental. This uplift was the region and are dated between the upper Miocene and promoted by tecto-volcanic activity in the region and was lower Pleistocene [7]. related to the opening of the Proto Gulf of California dur- ing the lower and middle Miocene [54,55]. This biogeo- The second event within the Central-Eastern clade was the graphical pattern is in agreement with findings made in isolation of A. amecae in the Ameca River from the ances- other freshwater fishes, such as goodeids (the separation tor of the remaining members of the Central-Eastern of the Characodontini tribe from the remaining tribes of clade, A. tincella and A. lacustris. This cladogenetic event the subfamily Goodeinae), and with the separation of the also took place in the upper Pliocene, approximately 2.4 two divergent groups of the poecilid Poeciliopsis spp., MYA (3.1-1.9 MYA). This pattern clearly indicates an which occurred about 15.5 MYA and between 8 and 16 ancient connection between the upper Ameca River and MYA, respectively [18,20]. the Lerma-Chapala River system, which was further dis- rupted by geological activity along the Ameca and San The separation of the three main clades of Algansea is Marcos faults, which occurred throughout the Pliocene dated at some point between the Miocene and Pliocene, (Figure 6) [58]. It is also related to the dry period reported about 6.2 MYA (7.8-4.7 MYA), a period recognised for a in the area between 3.5 to 1.8 MYA [56]. Ferrari et al. [60] series of profound global palaeoclimatic changes and mentioned the presence of volcano-lacustrine deposits high geological activity in central Mexico [56]. In particu- along these faults, which probably indicates that the for- lar, during the formation of the western part of the TMVB, mation of these depressions occurred during the early several events were responsible for the isolation of major Pliocene. These results are in agreement with findings in groups within Algansea and within other groups of fresh- other groups of freshwater fishes that currently occur in water fish (Figure 6). Including the genus Poeciliopsis [18], both basins, such as the cyprinids Yuriria amatlana, Zoog- the formation of the two main lineages within the oneticus tequila, and Ameca splendens (all endemic to the cyprinid Notropis of central Mexico [23], and the diversifi- Ameca River basin) and their sister groups, inhabitants of cation of three tribes among the Goodeinae [20], all of the Lerma-Chapala system [7,11]. which coincided with these events. The cladogenetic event of the species pair A. lacustris, Within the Western clade, the cladogenesis of A. avia and endemic to Patzcuaro Lake, and A. tincella, a widespread A. monticola, both inhabitants of the Santiago River, was species in the Lerma, Panuco, Balsas, and Verde de San- dated to the Pliocene, 4.5 MYA (6.2-3.2 MYA). Currently, tiago rivers basins and Cuitzeo Lake, is dated to the upper these species are separated by a deep canyon in the main Pliocene, 1.9 MYA (2.6-1.3 MYA). Although the forma- channel of the Santiago River; therefore, the formation of tion and evolution of Patzcuaro Lake remains unclear and this geological structure may be responsible for the sepa- controversy surrounds the age of the lake's origin, vol- ration of these species [57]. The origin of the Santiago canic activity in the Tarasco Corridor has been previously River Canyon is related to the Santa Rosa faulting, which proposed as one of the causes of diversification of other is dated to the early Pliocene [58] (Figure 6). The presence fish species in the region in the last 3 MYA [19]. of A. avia in the Compostela River in Nayarit can be attrib- uted to a river capture event frequently occurring in the The historical biogeography of the genus Algansea seems upper part of rivers in central México due to headwaters to be closely linked to the intense geological and climatic erosion. This event is dated to less than 1 MYA [20] and activity of central Mexico, which is one of the most com- coincides with the distribution pattern of the goodeid plex and active geological areas in the world [61-63]. Xenotoca eiseni. Based on both historical and geological data, most of the

Page 11 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

FigureSpeciation 6 events within the genus Algansea. Speciation events within the genus Algansea.

Page 12 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

cladogenetic events within the genus Algansea were the clear and shallow water (less than one meter deep). On result of vicariant events (Figure 6). This intense activity the other hand, within the member of the Central-Eastern has been proposed as the main cause of diversification in clade, A. lacustris, is restricted to lakes; A. amecae and A. other freshwater species of central Mexico, such as the barbata live in streams with slow to moderate water flow, Goodeids [7,20,64], Poeciliids [18,65], helminth para- sandy silt bottoms, and water depths reaching more than sites [66-70], snakes [71], frogs [72,73], and ambystoma- one meter. Furthermore, A. tincella occurs in a variety of tid salamanders [74]. habitats ranging from small streams to lakes. These eco- logical aspects are closely tied with larger body size and, Most of the aforementioned vicariant events are highly therefore, are found mainly in the members of the Cen- consistent with the topology obtained from the combined tral-Eastern clade. analyses. Still, more groups must be studied in the same context for a better understanding of the processes that led Although our study did not include the species A. popoche to the diversification of the complex biota that occurs in (endemic to Chapala Lake), we could assume, based on this transitional zone. the distribution of this species, that it would likely be included within the Central-Eastern clade. Barbour and Morphological evolution and taxonomic implications Miller [21] proposed A. popoche as the sister species of A. North American cyprinids show a diverse and complex lacustris. However, we were unable to collect specimens of morphological evolution; this complexity is probably the A. popoche because this species may have gone extinct, and main reason to consider theses fishes as one of the most we could not further corroborate this idea. However, most taxonomically difficult groups on the continent [26]. of the recognised species within the genus Algansea (A. aphanea, A. avia, A. barbata, A. lacustris, A. monticola, A. tin- According to our results, the barbeled species group (A. cella, [5], and A. amecae [24]) were corroborated by our monticola, A. aphanea and A. barbata) that was based solely phylogenetic analyses. Two species regarded as subspecies on morphology from previous phylogenetic hypotheses of A. monticola by Barbour and Miller [21], A. monticola of the genus Algansea [21,36] was not recovered as mono- avia and A. monticola monticola, exhibited genetic diver- phyletic; this result corroborates the conflictive role that gence levels in our study that allowed us to further test the the maxillary barbels have for the classification of the fam- hypothesis that A. avia is an independent species. These ily Cyprinidae [25,26]. However, even though maxillary tests granted species rank to A. avia and corroborated the barbels seem to have evolved independently several times morphological differentiation found by Jensen and Bar- [25], the presence of such structures in A. barbata is bour [36] and Barbour and Miller [57]. regarded as a plesiomorphic character for the clade to which this species belong. Instead, the lack of barbels in Conclusion the subgroup of the Central-Eastern clade, together with a The present study is based on two independent sources of higher number of gill rakers, six other character states evidence (mitochondrial and nuclear molecular markers) (except for body size smaller than 100 mm and the that corroborate the monophyly of the genus Algansea. rounded neurocranium dome), are explained as derived Despite the low resolution found in the phylogenies characters (Figure 5A-G). within this genus, three well-supported clades were recov- ered: the Southern clade, including A. aphanea; the West- In particular, possessing a larger body is an adaptation in ern clade, including A. avia and A. monticola; and the North American that inhabit open-water habi- Central-Eastern clade, including A. barbata, A. amecae, A. tats [75]. This suggests that the morphological traits found lacustris, and A. tincella. In particular, the monophyly of in the members of Algansea that constitute the Central- the Central-Eastern clade was found, and our results did Eastern clade are clearly associated with lacustrine envi- not support the groups proposed by Barbour and Miller ronments. This adaptation might have been promoted [21] based on the presence of barbels. Historical biogeog- when species existed in the paleolakes in central Mexico raphy patterns of this clade are shared with other groups (see Biogeographic implications). of co-distributed fishes in central Mexico. Additionally, a larger body size in members of the Central-Eastern clade Barbour and Miller [21] described some characteristics of suggests an evolutionary adaptation that arose when these the ecological conditions of the habitats where species of species colonised the lacustrine environments in which Algansea occur. These data, and our own observations in they occur in central Mexico. the sampling localities, allow us to conclude that the spe- cies forming the Western and Southern clades are We corroborate the western origin of the genus Algansea; restricted to the upper parts of rivers, which are character- however, the sister species is the monotypic Agosia. Most ised by swift currents with rocky and gravel bottoms and of the cladogenetic events are associated with vicariance

Page 13 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

occurring in the region as a result of tectonic activity and Big Dye Deoxy Terminator cycle-sequencing kit (Applied climatic changes. Such events seem to have also shaped Biosystems Inc.) in an ABI PRISM 3700 DNA analyser. For the evolutionary history of other freshwater fish groups. cyt b, chromatograms and alignments were visually checked and the S7 sequences were aligned using Clustal Methods × 1.83 with default parameters [79]. All sequences were Species and Sampling revised and verified using MEGA 3.1 [80]. GeneBank Fin clips were obtained from 48 specimens in 20 localities accession numbers are presented in Additional File 1. We (Additional File 1 and Figure 4) belonging to seven puta- obtained the complete mitochondrial cyt b (total of 1140 tive species of the genus Algansea (A. aphanea, A. avia, A. bp) of 47 specimens from 25 localities (Additional File 1). barbata, A. lacustris, A. monticola, A. tincella, and A. ame- Based on at least one representative specimen from each cae). Fishes were collected by electrofishing and seine mitochondrial clade, we sequenced the first intron of the nets. The only species not included in the analysis was A. ribosomal protein S7 (965 bp, including gaps) for a sub- popoche, an endemic species from Chapala Lake [5]. This set of 19 specimens from 16 localities (Additional File 1). species has not been found in the locality despite a great The cyt b matrix was increased with previously published sampling effort made by different research groups, raising sequences for Algansea species (Schönhuth et al. [10]: the possibility that the species is now extirpated or extinct DQ324088-DQ324092). [76]. Additionally, fin clips were collected from represent- ative species of the cyprinid genera Agosia, Campostoma, Phylogenetic analysis Dionda, and Gila (Table 1). To test the monophyly of the Homologous regions were aligned manually against pre- genus Algansea, and due to the relationships found viously published cyt b sequences of Algansea [10]. The between Algansea and other North American cyprinids nucleotide composition and base frequencies were exam- [10,12], the first three genera were included as ingroups, ined, and the homogeneity test of base frequencies (for leaving Gila as an outgroup. Voucher specimens of all spe- each single codon position and all the positions) was car- cies were deposited in the Museo Nacional de Ciencias ried out for all taxa and both genes with the program Naturales (MNCN), Madrid, Spain, and at the Colección PAUP 4.1 [81]. Furthermore, the saturation of transition de Peces de la Universidad Michoacana de San Nicolas de and transversion changes was checked for each gene by Hidalgo (CPUM), Michoacán, México. plotting the absolute number of changes of each codon position against their patristic distances (p). There was no DNA extraction, PCR and sequencing evidence of saturation for any of the sequence data sets Total genomic DNA was extracted from ethanol-preserved (data not shown). fin clips according to standard CTAB and phenol-chloro- form extraction procedures [77]. Both the complete Phylogenetic analyses were carried out for both genes sep- sequence of cyt b and the first intron of ribosomal protein arately and as a combined data set using Maximum Parsi- gene S7 were amplified via polymerase chain reaction mony (MP) and Bayesian Inference (BI) analyses. The MP (PCR). For both genes, reaction amplifications were car- analyses were performed in PAUP* version 4.1 using heu- ried out in 25 μl reactions containing: 2.5 μl 10× buffer ristic searches (TBR Branch swapping; MULPARS option with MgCl2 (Biotools), 0.5 μl of each dNTP (10 μM), 0.3 in effect) with 10 random stepwise additions of taxa. μl of each primer, 1-2 μl genomic DNA (50 ng/Ml), 1 unit Analyses of the S7 gene and combined matrix were done of Taq DNA polymerase (Biotools), and distilled water to including or excluding the indels. bring the final reaction volume to 25 μl. For the BI analyses, the best-fit model for the different For the cyt b gene, the primers used were those primers in genes and each codon position (for cyt b in particular) Machordom and Doadrio [78]. For the first intron of S7, were selected using Modeltest 3.7 [32] based on the Baye- we used the primers of Chow and Hazama [28]. The sian Information Criterion (BIC). Bayesian analyses were amplification was performed using the following condi- performed in MrBayes 3.1.2 [82], using two independent tions. For cyt b, 35 cycles were used: denaturation at 94°C runs of four Metropolis-Coupled Markov Chain Monte for 45 sec, annealing at 46°C for 1 min, and extension at Carlo (MCMC) of 1,000,000 generations each to estimate 72°C for 1:30 min. A final extension at 72°C for 5 min the posterior probability distribution. The combined was performed to completely extend the amplified prod- sequence matrices were partitioned per gene fragment, uct. For the S7 first intron, 30 cycles were performed using and independent model parameters were estimated for a denaturation at 94°C for 1 min, annealing at 54°C for each partition. Topologies were sampled every 100 gener- 1:30 min, and an extension at 72°C for 2 min; then, a ations. Once the average standard deviation of split fre- final extension was performed at 72°C for 7 min. quencies was less than 0.01, as suggested by MrBayes 3.1.2 [82], convergence between runs was checked. This was PCR products were purified with the QIAquick (QIAGEN) accomplished by comparing the 50% majority rule con- kit, checked on 1% agarose gels, and sequenced using the sensus tree for each run (after discarding the first 20,000,

Page 14 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

25,000 and 10,000 generations for the cyt b, S7, and com- removing 10% of the generations from each analysis as bined gene data sets, respectively), and no incongruence "burn-in", the ESS exceeded 210 for all parameters. between the runs were found. Authors' contributions The robustness of the clades was assessed using bootstrap- RP-R collected the samples, carried out the molecular ping (1,000 pseudoreplicates) for the MP analyses and work, analysed the data, and drafted the original manu- Bayesian posterior probabilities for the BI analyses. script. GP-PdL participated in the study design, contrib- uted to the draft of the original manuscript, and The ancestral character-state reconstruction of the evolu- contributed to the improvement of all versions of the tion of some of the morphological characters consisted of manuscript. OD-D conceived the study, collected the sam- a subset of nine characters from a data matrix of twenty- ples, help with the analyses, and contributed to the six characters used in Barbour and Miller's [21] and Jensen improvement of all versions of the manuscript. ID con- and Barbour's [36] phylogenies (Additional File 2). These ceived the study, participated in its design and coordina- characters were selected because they were synapomor- tion, and contributed to the improvement of all versions phies supporting particular clades in those studies. The of the manuscript. All authors read and approved the final software MacClade version 3.06 [83] was used to recon- manuscript. struct character evolution. Additional material Molecular clock and divergences The uncorrected p distances and absolute changes were calculated using all the specimens analysed. The average p Additional file 1 Localities and Genebank accessions numbers of individuals from the distances between the main clades were also calculated for species analysed. the table provided specific information about of sampled both markers using the program MEGA v.3.1 [80]. localities, the number of individuals analysed for cytochrome b and S7 intron 1, and the Genebank accessions numbers. A relaxed molecular clock Bayesian approach using cyt b Click here for file gene (1140 bp) was performed to infer the time to the [http://www.biomedcentral.com/content/supplementary/1471- most recent common ancestor (TMRCA) in the different 2148-9-223-S1.pdf] species and groups within Algansea and with respect to Additional file 2 Agosia. Because of the lack of reliable fossil records for the Character-state matrix. the table consists in a matrix of character state- genus Algansea, a molecular clock of 1.05% per million coded values for species of the genus Algansea and the sister group Agosia year, which has been established for cyt b in North Amer- chrysogaster. All character state-coded values for Algansea, except the ican Phoxinini by Dowling et al. [84] and widely applied fourth and fifth characters, were based on Barbour and Miller (1978). to Euro Asiatic cyprinids [85-89], was used. The molecular Gut flexure, supraethmoid orientation, and neurocranium dome were clock was used to i) estimate the divergence times of the based on Jensen and Barbour (1981). For standard length, we followed the criteria of Barbour and Miller (1978), and length was recoded for the main cladogenetic events implicated in the origin and present study. diversification of Algansea and ii) test if these estimates are Click here for file in agreement with the historical geologic events in the [http://www.biomedcentral.com/content/supplementary/1471- central Mexico region that may be responsible for the ori- 2148-9-223-S2.pdf] gin and diversification of other co-distributed groups of fishes. Divergence times and their confidence intervals were estimated using a relaxed clock model in BEAST v1.4.6 [90], with a strategy that included branch rates Acknowledgements drawn from an uncorrelated log-normal distribution [91]. The authors wish to thank the Tiacaque fish farm workers for providing the The estimates of divergence times was done with a Yule A. barbata specimen. We also thank the fisherman from the Uranden natural reserve and Patzcuaro Lake for providing specimens of A. lacustris. We tree prior, the branch length substitution rate sampled thank Rogelio Rosas-Valdez and Jaquelina Bravo-Arteaga for their help dur- from a prior normal distribution (with a mean value of ing field work. We thank H. Gante for your revisions and comments to this 0.010 and a standard deviation of 0.001 [92]) and work. This study was partially funded by grants from CGL2006-12325/BOS, included just one representative of each species. The the program PAPIIT-UNAM-IN209608, and CONACYT (grant number TrN+G substitution model was used, and three MCMC for 83043) to GPPDL. RPR thank the Consejo Nacional de Ciencia y Tec- 90 × 106 generations were run. The remaining parameters nología for the scholarship. were set as default options and changed as recommended by the BEAST output file. We checked for the effective References 1. Morrone JJ: Hacia una síntesis biogeográfica de México. Revista sample size (ESS) convergence and the stationary of the Mexicana de Biodiversidad 2005, 76:207-252. different analyses in Tracer 1.4 [92] and combined the 2. Barbour CD: The systematics and evolution of the genus Chi- results in the BEAST module LogCombiner 1.4.4. After rostoma Swainson (Pisces, Atherinidae). Tulane Studies in Zool- ogy and Botanical 1973, 18:97-141.

Page 15 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

3. Barbour CD: A biogeographical history of Chirostoma (Pisces: 23. Schönhuth MS, Doadrio I: Phylogenetic relationships of Mexican Atherinidae): A species flock from the Mexican Plateau. minnows of the genus Notropis (: Cyprinidae). Copeia 1973, 1973:533-556. Biological Journal of the Linnean Society 2003, 80:323-337. 4. Guzmán-Arroyo F: Osteología y variación no geográfica de la 24. Pérez-Rodríguez R, Pérez-PoncedeLeón G, Domínguez-Domínguez suspensión de la aleta anal de Goodea luitpoldi, (Osteich- O, Doadrio I: A new species of the genus Algansea Girard, thyes:). Universidad, Ciencia y Tecnología 1994, 3:33-41. 1856 (Actinopterygii: Cyprinidae) from the Ameca River 5. Miller RR, Minkley WL, Norris SM: Freshwater fishes of Mexico. basin, in Central Mexico. Revista Mexicana de Biodiversidad 8:340. 1st edition. Chicago: University of Chicago Press; 2005. 25. Howes GJ: Systematics and biogeography: an overview. In 6. Miller RR, Smith ML: Origin and geography of the fishes of cen- Cyprinids fishes: Systematics, biology and exploitation Edited by: Winfield tral Mexico. In The Zoogeography of North American Freshwater Fishes IJ, Nelson JS. London: Chapman & Hall; 1991:1-33. Edited by: Hocutt CH, Wiley EO. New York: Wiley-Interscience Pub- 26. Simons AM, Berendzen PB, Mayden RL: Molecular systematics of lications; 1986:487-519. North American phoxinin genera (Actinopterygii: Cyprini- 7. Domínguez-Domíguez O, Doadrio I, Pérez-PoncedeLeon G: Histor- dae) inferred from mitochondrial 12S and 16S ribosomal ical biogeography of some river basins in central Mexico evi- RNA sequences. Zoological Journal of the Linnean Society 2003, denced by their goodeine freshwater fishes: a preliminary 139:63-80. hypothesis using secondary Brooks parsimony analysis. Jour- 27. Lydeard C, Roe K: The phylogenetic utility of the mitochon- nal of Biogeography 2006, 33:1437-1447. drial cytochrome b gene for inferring relationships among 8. Espinosa H, Gaspar M, Fuentes P: Listados faunísticos de México. III. Los Actinopterigian fishes. In Molecular systematics of fishes Edited by: peces dulceacuícolas mexicanos México DF: Instituto de biología, Uni- Kocher TD, Stepien CA. New York Academic Press; 1997:285-302. versidad Nacional Autónoma de México; 1993. 28. Chow S, Hazama K: Universal PCR primers for S7 ribosomal 9. Doadrio I, Domínguez O: Phylogenetic relationships within the protein gene introns in fish. Molecular Ecology 1998, 7:1247-1263. fish family Goodeidae based on cytochrome b sequence 29. Farias IP, Ortí G, Sampaio I, Schneider H, Meyer A: The cyto- data. Molecular Phylogenetic and Evolution 2004, 31:416-430. chrome b gene as a phylogenetic marker: the limits of reso- 10. Schönhuth S, Doadrio I, Mayden R: A biogeographic perspective lution for analyzing relationships among cichlid fishes. Journal on the phylogeny of mexican cyprinids (Actinopterygii: of Molecular Evolution 2001, 53:89-103. Cyprinidae). In Studies of north American desert fishes: in honor of E P 30. Bernardi G, Bucciarelli G, Costagliola D, Robertson DR, Heiser JB: (Phil) Pister, conservationist 1st edition. Edited by: Lozano-Vilano ML, Evolution of coral reef fish Thalassoma spp. (Labridae). 1. Contreras-Balderas S. Monterrey, NL: Universidad Autónoma de Molecular phylogeny and biogeography. Marine Biology 2004, Nuevo León; 2006:102-124. 144:369-375. 11. Domínguez-Domínguez A, Pompa-Domínguez A, Doadrio I: A new 31. Johnson JB, Dowling TE, Belk MC: Neglected of rare species of the Genus Yuriria Jordan & Evermann, 1896 desert fishes: congruent evidence for two species of leather- (actinopterygii, cyprinidae) from the Ameca Basin if the side chub. Systematic Zoology 2004, 53:841-855. Central Mexican Plateau. Graellsia 2007, 63:259-271. 32. Posada D, Crandall KA: Modeltest: testing the model of DNA 12. Schönhuth S, Doadrio I, Domínguez-Domínguez O, Hillis DM, substitution. Bioinformatics 1998, 14:817-818. Mayden R: Molecular evolution of southern North American 33. Bielawski JP, Gold JR: Phylogenetic Relationships of Cyprinid Cyprinidae (Actinopterygii), with the description of the new Fishes in Subgenus Notropis Inferred from Nucleotide genus Tampichthys from central Mexico. Molecular Phylogenet- Sequences of the Mitochondrially Encoded Cytochrome b ics and Evolution 2008, 47:729-756. Gene. Copeia 2001:656-667. 13. De Buen F: Los Lagos Michoacanos I. Caracteres generales. El 34. Orti G, Petry P, Porto JIR, Jégu M, Meyer A: Patterns of nucleotide Lago de Zirahuén. Revista de la Sociedad Mexicana de Historia Nat- change in mitochondrial ribosomal RNA genes and the phy- ural 1943, 4:211-232. logeny of piranhas. Journal of Molecular Evolution 1996, 42:169-182. 14. Álvarez del Villar J: Ictiología Michoacana V. Origen y dis- 35. Lavoué S, Sullivan JP, Hopkins CD: Phylogenetic utility of the first tribución de la ictiofauna dulceacuícola Michoacana. Anales de two introns of the S7 ribosomal protein gene in african elec- la Escuela Nacional de Ciencias Biológicas 1972, 19:155-161. tric fishes (Mormyroidea: Teleostei) and congruence with 15. Parenti L: A phylogenetic and biogeographic analysis of other molecular markers. Biological Journal of the Linnean Society cyprinidontiform fishes (Teleostei, Atherinomorpha). Bulletin 2003, 78:273-292. of the American Museum of Natural History 1981, 168:335-557. 36. Jensen RJ, Barbour CD: A phylogenetic reconstruction of the 16. Echelle AA, Echelle AF: Evolutionary genetics of a "species mexican cyprinid fish genus Algansea. Systematic Zoology 1981, flock": atherinid fishes on the Mesa Central of Mexico. In Evo- 30:41-57. lution of fish species flocks Edited by: Echelle AA, Kornfield I. Orono: 37. Minckley WL, Hendrickson DA, Bond CE: Geography of western University of Maine Press; 1984:93-110. North American freshwater fishes: description and relation- 17. Moncayo-Estrada R, Israde-Alcantara I, Garduño-Monroy VH: La ships to intracontinental tectonism. In Zoogeography of North cherehuita Hubbsina turneri De Buen (1941) (Pisces, Good- American Freshwater fishes Edited by: Hocutt CH, Wiley EO. New eidae): origen, distribución y su uso en la regionalización de York: Wiley-Interscience Publications; 1986:519-613. la cuenca del lerma. Hidrobiológica 2001, 11:1-13. 38. Sullivan JP, Lavoué S, Hopkins CD: Discovery and phylogenetic 18. Mateos M, Sanjur OI, Vrijenhoek C: Historical biogeography of analysis of a riverine species flock of african electric fishes the livebearing fish genus Poeciliopsis (Poecilidae: Cyprino- (Mormyridae: Teleostei). Evolution 2002, 56:597-616. dontiformes). Evolution 2002, 56:972-984. 39. Slowinski JB: Molecular politomies. Molecular Phylogenetics and 19. Domínguez-Domínguez O, Alda F, Pérez-PoncedeLeón G, García- Evolution 2001, 19:114-120. Garitagoitia JL, Doadrio I: Evolutionary history of the endan- 40. Goncalves H, Martínez-Solano I, Ferrand N, García-París M: Con- gered fish Zoogoneticus quitzeoensis (Bean, 1898) (Cyprino- flicting phylogenetic signal of nuclear vs mitocondrial DNA dontiformes: Goodeidae) using a sequential approach to markers in midwife toads (Anura, Discoglossidae, Alytes): phylogeography based on mitochondrial and nuclear DNA Depp coalescence or ancestral hybridization? Molecular Phylo- data. BMC Evololutionary Biology 2008, 8:161. genetics and Evolution 2007, 44:494-500. 20. Domínguez-Domínguez O, Pedraza-Lara CP, Gurrola-Sánchez N, 41. Holland BR, Benthin S, Lockhart PJ, Moulton V, Huber KT: Using Pérez-Rodríguez R, Alcaraz L, Perea S, Ornelas CP, Israde-Alcántara supernetworks to distinguish hybridization from lineage- I, Garduño-Monroy VH, Doadrio I, Pérez-PoncedeLeón G, Brooks sorting. BMC Evolutionary Biology 2008, 8:202. DR: Historical biogeography of the Goodeinae (Cyprinodon- 42. Annilo T, Jelina J, Pata I, Metspalu : Isolation and characterization tiforms). In Viviparous fishes II Edited by: Uribe-Aranzabal MC, Grier of the Mouse ribosomal proteína S7 gene. Biochemistry and H. Florida: New Life Publications in press. Molecular Biology International 1998, 46:287-295. 21. Barbour CD, Miller RR: A revision of the genus Algansea. Mis- 43. Chow S, Scholey VP, Nakazawa A, Margulies JB, Wexler JB, Olson RJ, cellaneous Publications Museum of Zoology, University of Michigan 1978, Hazama K: Direct Evidence for Mendelian Inheritance of the 155:1-72. Variations in the Ribosomal Protein Gene Introns inYellow- 22. Lyons J, Polaco OJ, Cochran PA: Morphological variations among fin Tuna (Thunnus albacares). Marine Biotechnology 2001, the Mexican lampreys (Petromyzontidae: Lampetra: subge- 3:22-26. nus Tetrapleurodon). Southwestern Naturalists 1996, 41:365-374.

Page 16 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

44. He S, Mayden RL, Wang X, Wang W, LTang K, Chen WJ, Chen Y: ras, México. Boletín de la Sociedad Geológica Mexicana 2000, Molecular phylogenetics of the family Cyprinidae (Actinop- 53:57-59. terygii: Cypriniformes) as evidenced by sequence variation 64. Gesundheit P, Macias-García C: Biogeografía cladísta de la in the first intron of S7 ribosomal protein-coding gene: Fur- familia goodeidae (). In Una prespectiva lat- ther evidence from a nuclear gene of the systematic chaos in inoamericana de la biogeografía Edited by: Llorente J. México DF: the family. Molecular Phylogenetics and Evolution 2008, 46:818-829. Comisión Nacional para el Conocimiento y Uso de la Biodiverisidad; 45. Holder MT, Anderson JA, Holloway AK: Difficulties in detecting 2006:319-337. hybridization. Systematic Biology 2001, 50:978-982. 65. Mateos M: Comparative phylogeography of livebearig fishes in 46. Buckley TR, Cordeiro M, Marshall DC, Simon C: Differentiating the genera Poeciliopsis and Poecilia (Poeciliidae: Cyprino- between hypotheses of lineage sorting and introgression in dontiformes) in central Mexico. Journal of Biogeography 2005, New Zealand alpine cicadas (Maoricicada Dugdale). System- 32:775-780. atic Biology 2006, 55:411-425. 66. Pérez-PoncedeLeón G: Biodiversity and biogeographic patterns 47. Miller DL, Behnke RJ: Two new intergeneric cyprinid hybrids in the Mesa Central of México: insights from host-parasities from the Bonneville Basin, Utah. Copeia 1985, 1985:509-515. systems. Journal of Parasitology 2003, 89:126-133. 48. Dowling TE, Smith GR, Brown WM: Reproductive isolation and 67. Aguilar-Aguilar R, Contreras-Medina R, Salgado-Maldonado G: Parsi- introgression between Notropis cornutus and Notropis mony analysis of endemicity (PAE) of Mexican hydrological chrysocephalus (family Cyprinidae): comparson of morphol- basins based on helminth parasites of freshwater fishes. Jour- ogy allozymes, and mitochondrial DNA. Evolution 1989, nal of Biogeography 2003, 30:1861-1872. 43:620-634. 68. Pérez-PoncedeLeón G: processes AC BohpoffiMtsfpa: Biogeog- 49. DeMarias BD, Dowling TE, Douglas ME, Minckley WL, Marsh PC: raphy of helminth parasites of freshwater fishes in Mexico: Origin of Gila seminuda (Teleostei:Cyprinidae) through the search for patterns and processes. Journal of Biogeography introgressive hybridization: Implications for evolution and 2005, 32:645-659. conservation. Proceedings of the National Academy of Sciences of the 69. Mejia-Madrid HM, Vázquez-Domínguez E, Pérez-PoncedeLeón G: United States of America 1992, 89:2747-2751. Phylogeography and freshwater basins in central Mexico: 50. Dowling TE, DeMarias BD: Evolution significance of introgres- recent history as revealed by the fish parasite Rhabdochona sive hybridization in cyprinid fishes. Nature 1993, 362:444-446. lichtenfelsi (Nematoda). Journal of Biogeography 2007, 51. Bostrom MA, Collete BB, Luckhurst BE, Reece KS, Graves JE: 34:787-801. Hybridization between two serranids, the coney (Cephalo- 70. Rosas-Valdéz R, Domíngue-Domínguez O, Choudhury A, Pérez-Pon- pholis fulva) and the creole-fish (Paranthias furcifer), at Ber- cedeLeón G: Helminth parasites of the Balsas catfish Ictalurus muda. Fishery Bulletin 2002, 100:651-661. balsanus in several localities of the Balsas River Drainage, 52. Rosenthal GG, Rosa-Reyna XF, Kazians S, Stephens MJ, Morizot DC, Mexico: Species composition and biogeographical affinities. Ryan MJ, León JGd: Dissolution of sexual signal complexes in a Comparative parasitology 2007, 74:204-210. hybrid zone between the swordtails Xiphophorus birch- 71. Contant R: Observations on Garter Snakes of the Thamno- manni and Xiphophorus malinche (Poeciliidae). Copeia 2003, phis eques Complex in the Lakes of Mexico's Transvolcanic 2003:299-307. Belt, with Descriptions of New Taxa. American Museum Novi- 53. Moore WS: Inferring phylogenies from mtDNA variation: tates, American Museum of Natural Histroy, NY 2003, 3406:1-64. mitochondrial-gene trees versus nuclear-gene trees. Evolution 72. Mulcahy DG, Mendelson JR III: Phylogeography and speciation of 1995, 49:718-726. the morphologically variable, widespread species Bufo valli- 54. Henry CD, Aranda-Gómez JJ: Plate interactions control middle- ceps, based on molecular evidence from mtDNA. Molecular late Miocene, proto-Gulf and Basin and Range extension in Phylogenetics and Evolution 2000, 17:173-189. the southern Basin and Range. Tectonophysics 2000, 318:1-26. 73. Zaldívar-Riverón A, León-Regagnon V, Nieto-MontesdeOca A: Phy- 55. Ferrari L, López-Martínez M, Rosas-Elguera J: Ingnimbrite flare-up logeny of the Mexican coastal leopard frogs of the Rana ber- and deformation in the southern Sierra Madre Occidental, landieri group based on mtDNA sequences. Molecular western México. Implications for the late subduction history Phylogenetics and Evolution 2004, 30:38-49. of Farallon plate. Tectonics 2002, 21:1-24. 74. Weisrock DW, Shaffer HB, Storz BL, Storz SR, Voss SR: Multiple 56. Israde-Alcántara I, Wade M, Garduño-Monroy VH, Barron J: Estrati- nuclear gene sequences identify pfulogenetic species bound- grafía y encuadramiento geodinámico de las cuencas lacus- aries in the rapidly radiating clade of Mexican ambystomatid tres del centro de México. Unión Mexicana de estudios del salamanders. Molecular Ecology 2006, 15:2489-2503. Cuaternario Fondo de Cultura Económica in press. 75. Chan MD: Fish ecomorphology: predicting habitat prefer- 57. Barbour CD, Miller RR: Diversification in the mexican cyprinid ences of stream fishes from their body shape. In PhD thesis Fac- fish Algansea monticola (Pisces: Cyprinidae), with descrip- ulty of the Virginia Polytechnic Institute and the State University; tion of a new subspecies. Copeia 1994:662-676. 2001. 58. Ferrari L, Rosas-Elguera J: Late Miocene to Quaternary exten- 76. Lyons J, González-Hernández G, Soto-Galera E, Guzmán-Arroyo M: sion an the northern boundary of the block, western Decline of freshwater fishes and fisheries in selected drain- Mexico: The Tepic-Zacoalco rift revised. Geological Society of ages of west-central Mexico. Fisheries 1998, 23:10-18. America Special Paper 1999, 334:1-23. 77. Sambrook J, Fritsch EF, Maniatis T: Molecular cloning: a laboratory man- 59. Israde-Alcántara I, Garduño-Monroy VH: Lacustrine record in a ual New York: Cold Spring Harbor Laboratory Press; 1989. volcanic intra-arc setting: the evolution of Late Neogene 78. Machordom A, Doadrio I: Evidence of a Cenozoic Betic-Kabilian Cuitzeo basin system (central-western Mexico, Michoacan). connection based on freshwater fish phylogeography (Lucio- Palaeogeography, Palaeoclimatology and Palaeoecology 1999, barbus, Cyprinidae). Molecular Phylogenetics and Evolution 2001, 151:209-227. 18:252-263. 60. Ferrari L, López-Martínez M, Aguirre-Díaz G, Carrasco-Núñez G: 79. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG: The Space-time patterns of Cenozoic arc volcanism in central ClustalX windows interface: flexible strategies for multiple Mexico: from the sierra madre occidental to the Mexican sequence alignment aided by quality analysis tools. Nucleic Volcanic Belt. Geology 1999, 27:303-306. Acids Researh 1997, 24:4876-4882. 61. Tamayo LT, West RC: The hydrology of Middle America. In 80. Kumar S, Tamura K, Nei M: MEGA3: Integrated software for Handbook of Middle America indians Edited by: West RC. Texas: Uni- molecular evolutionary genetics analysis and sequence align- versity of Texas Press; 1964:84-121. ment. Briefings in Bioinformatics 2004, 5:150-163. 62. Ferrusquia-Villafranca I: Geología de México: una synopsis. In 81. Swofford DL: Phylogenetic analysis using parsimony (* and other methods) Diversisdad Biológica de México: orígenes y distribución Edited by: Rma- V 4.0b 10 Sunderland MA, Sinauer Associates; 2004. moorthy TP, Bye R, Lot A, Fa J. México DF: Instituto de biología, Uni- 82. Huelsenbeck JP, Ronquist F: MRBAYES: Bayesian inference of versidad Nacional Autónoma de México; 1998:1-107. phylogeny. Bioinformatics 2001, 17:754-755. 63. Aranda-Gómez JJ, Henry CD, Luhr JF: Evolución tectomagmática 83. Maddison DR, WPMeC: McClade. Analysis of phylogeny and character post-paleocénica de la Sierra Madre Occidental y de la por- evolution v.4.05 OS X Sunderland MA, Sinauer Associates; 2002. ción meriodonal de la provincia tectónica de Cuencas y Sier-

Page 17 of 18 (page number not for citation purposes) BMC Evolutionary Biology 2009, 9:223 http://www.biomedcentral.com/1471-2148/9/223

84. Dowling TE, Tibbets CA, Minckley WL, Smith GR: Evolutionary relatioships of the plagopterins (Teleostei:Cyprinidae) from cytochrome b sequences. Copeia 2002, 2002:665-678. 85. Doadrio I, Carmona JA: Testing freshwater Lago Mare dispersal theory on the phylogeny relationships of Iberiancyprinid genera, Chondrostoma and Squalius. Graellsia 2003, 59:457-473. 86. Durand JD, Bianco PG, Laroche J, Gilles A: Insight into the origin of endemic Mediterranean ichthyofauna - Phylogeography of Chondrostoma genus (Teleostean, Cyprinidae). Journal of Heredity 2003, 94:315-328. 87. Doadrio I, Carmona JA: Phylogenetic relationships of the genus Chondrostoma using cytochrome b sequences. Molecular Phy- logenetics and Evolution 2004, 13:2807-2817. 88. Robalo JI, Sousa-Santos CS, Carvalho-Almada V, Doadrio I: Paleobi- ogeography of Two Iberian Endemic Cyprinid Fishes (Chon- drostoma arcasii-Chondrostoma macrolepidotus)Inferred from Mitochondrial DNA Sequence Data. Journal of Heredity 2006, 97:143-149. 89. Robalo JI, Carvalho-Almada V, Levya A, Doadrio I: Re-examination and phylogeny of the genus Chondrostoma based on mito- chondrial and nuclear data and the definition of 5 new gen- era. Molecular Phylogenetics and Evolution 2006, 42:362-372. 90. Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary anal- ysis by sampling trees. BMC Evolutionary Biology 2007, 7:214. 91. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A: Relaxed phyloge- netics and dating with confidence. PLoS Biology 2006, 4:e88. 92. Rambaut A, Drummond AJ: Tracer v1.4. 2007 [http:// beast.bio.ed.ac.uk/Tracer].

Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright

Submit your manuscript here: BioMedcentral http://www.biomedcentral.com/info/publishing_adv.asp

Page 18 of 18 (page number not for citation purposes)